面向低延时实时视频的多维跨层带宽预测

Multidimensional Cross-Layer Bandwidth Prediction for Low-LatencyReal-Time Video

5G网络的可用带宽是影响实时视频业务的关键要素之一,但如何在低延时实时视频业务下完成准确预测依旧是个难题。传统可用带宽预测算法通常依靠业务层的数据指标,根据发包策略完成预测。但此类算法在网络频繁变化的复杂场景下会出现预测滞后问题,严重影响了用户的接收视频质量。为解决此问题,文中提出了一种基于跨层多维参数的可用带宽预测算法,该算法综合考虑了业务层、物理层、网络层等的相关数据指标,并通过多个维度参数提升无线网络带宽探测的准确性。文中采用深度强化学习作为模型框架,针对不同的运动场景,通过跨层多维度的数据模型学习实现离线预测和在线预测的融合。同时,将网络丢包率、图像质量评估和端到端时延等链路影响因素作为约束条件,以实现预测模型在传输过程中的实时调整和优化。在半物理平台上的实验结果表明:文中所提算法的预测性能优于传统的预测算法,预测曲线与实际曲线的拟合程度高达95.8%以上;相比于单层预测算法,所提算法在步行、驾驶场景下的丢包率分别降低了47.3%和30.9%,接收视频质量提升了12%。

The available bandwidth of 5G networks is a pivotal factor influencing real-time video services. However, howto achieve accurate predictions in the context of low-latency real-time video services remains a difficult problem.Conventional algorithms for predicting available bandwidth typically rely on data metrics at the application layer andcomplete the forecast according to the packet transmission strategies. Such algorithms can lead to prediction lag in complexscenarios, thereby significantly impairing the received video quality for users. To address this concern, this study proposeda novel available bandwidth prediction algorithm based on cross-layer multi-dimensional parameters. This algorithmcomprehensively integrates pertinent data metrics from the application, physical, and network layers, utilizes multipledimensions of parameters to enhance the precision of wireless network bandwidth detection. In this paper, deepreinforcement learning was adopted as the model framework to integrate offline prediction and online prediction throughcross-layer and multi-dimensional data model learning for different motion scenarios. Furthermore, network packet lossrates, image quality assessments, end-to-end delays, and other link-related factors were introduced as constraints, to realizethe real-time adjustment and optimization of the prediction model during the transmission process. Experimental resultsconducted on a semi-physical platform show that: the prediction performance of the proposed algorithm is better than thetraditional prediction method, and the fitting degree of the prediction curve and the actual curve is more than 95. 8%;compared with the single-layer prediction algorithm, the packet loss rate of the proposed algorithm in walking and drivingscenarios decreases by 47. 3% and 30. 9%, respectively, and the quality of received video is improved by 12%.